Is there any software that can remove an accent or dialect from someone’s voice, or add one? This assumes that there’s a non-accent standard.
Well, everyone has an accent of some kind, even if it is “standard” and unremarkable. Are you talking about software that could take someone from Australia and make them sound like they are from New Jersey?
This is why I added “This assumes that there’s a non-accent standard.” But yes, also changing one dialect into another. I understand that the difference is primarily with vowels and diphthongs, so is that so hard to do?
There is software that translates to another language, but renders the speech using phonemes from the speaker’s own voice - so it sounds like you, speaking Japanese.
Accent though, I think would be more tricky, because it’s a lower level component of how people sound. The task almost summarises to “make it sound like me, but as if I sounded different”.
It’s more than vowels and diphthongs. Accent frequently involves consonantal sounds, which is why, e.g., it take a great deal of effort and close attention for a British English speaker to produce a plausible Irish accent - Hiberno-English makes consontantal distinctions which British English does not.
Plus, it’s also a matter of tempo and stress.
Plus, accent, dialect. idiom and vocabulary are closely intertwined. Currently the Straight Dope front page links to a 1987 piece of Cecil’s about the official residence of the British Prime Minister, which Cecil persistently refers to as “10 Downing”. You could read the piece in the most impeccable Received Standard accent, but that idiom marks it irrefutably as not the work of a user of British English. Idiomatic and vocabulary markers like this abound.
So, is there softwear to adjust pronunciation in the way you suggest? I don’t know that there is. It would be pretty complex to write and, no matter how well it was written, it would probably still not produce a plausible simulacrum of a speaker of a different variety of English.
I suppose you can already do this with existing speech-to-text and text-to-speech software, although that software is far from perfect. So convert your speech to text, then input it to a speech synthesiser, many of which can do various accents. Example: http://www.oddcast.com/home/demos/tts/tts_example.php .
Can you give a few examples?
Here is one example. Americans pronounce “ladder” and “latter” with the same consonant (a light tap on the roof of the mouth, or not even quite a tap), but with a noticeable difference in the length of the first vowel, while Brits will pronounce the latter with a distinct t. Similar remarks apply to “rider” and “writer”.
A bit off topic, but I know that the “t” is articulated in my dialect on the ridge behind the upper teeth, while in standard French it is actually on the back of the teeth. It wouldn’t at all surprise me to find dialects of English that use the French t.
When I was in Australia an American ex-pat told me a joke about a nurse asking a patient if he came to the hospital to die. The answer was “No, I came yesterdie.”
British English vs. Irish English was my question, actually.
Far be it from me
To fault a synecdoche,
But it wasn’t a rhetorical trope
Merely a place for Straight Dope
Yes, I understand now that consonants play a big part, e.g. the letter R. Compare Middle American, New England, British, Spanish, French, etc., etc. Very different sounds for only one letter. But I still don’t understand: If I hear a French person speaking English, there are certain sounds that I can easily identify as French. Why can’t a computer be taught to recognize the same sounds, and replace them with a different dialect?